smoother.visualization

Visualization functions for smoother.models.deconv

Functions

plot_celltype_props(p_inf, coords[, cell_type_names, ...])

Plot the deconvolution results.

clr_stable(props[, epsilon])

Apply centre log ratio transform (clr) to transform proportions to the real space.

ilr_stable(props[, epsilon])

Apply isometric log ratio transformation (ilr) to transform proportions to the real space.

cluster_features(→ pandas.Series)

Leiden clustering on the input features.

_get_cost_matrix(clu1, clu2)

Helper function to get the cost matrix for aligning two clusterings.

align_clusters(clu_list[, ref_ind])

Align the clusterings in clu_list to the reference clustering.

plot_spatial_clusters(clu_aligned_list, coords[, ...])

Plot clusters.

cluster_and_plot_celltype_props(p_inf_list, coords[, ...])

Cluster the cell-type proportions and visualize the results.

Module Contents

smoother.visualization.plot_celltype_props(p_inf, coords, cell_type_names=None, n_col=4, figsize=None)

Plot the deconvolution results.

Parameters:

  • p_inf – n_spots x n_groups. The inferred cell-type proportions.

  • coords – n_spots x 2. The coordinates of the spots.

  • cell_type_names – list of str. The names of the cell types.

  • n_col – int. The number of columns in the figure.

  • figsize – tuple. The size of the figure.

smoother.visualization.clr_stable(props, epsilon=1e-08)

Apply centre log ratio transform (clr) to transform proportions to the real space.

Parameters:

props – n_spots x n_groups. Rowsum equals to 1 or 0. If 0, the transformed vector will also be the zero vector.

smoother.visualization.ilr_stable(props, epsilon=1e-08)

Apply isometric log ratio transformation (ilr) to transform proportions to the real space.

smoother.visualization.cluster_features(features, transform='pca', n_neighbors=15, res=1) → pandas.Series

Leiden clustering on the input features.

smoother.visualization._get_cost_matrix(clu1, clu2)

Helper function to get the cost matrix for aligning two clusterings.

smoother.visualization.align_clusters(clu_list, ref_ind=None)

Align the clusterings in clu_list to the reference clustering.

Parameters:

  • clu_list – list of clustering result.

  • ref_ind – int. The index of the reference clustering. If None, use the clustering with the largest number of clusters.

smoother.visualization.plot_spatial_clusters(clu_aligned_list, coords, names=None, n_col=4)

Plot clusters.

Parameters:

  • clu_aligned_list – list of aligned clusterings.

  • coords – n_spots x 2. Coordinates of the spots.

  • names – list of str. Names of the deconvolution model.

  • n_col – int. Number of columns in the plot.

smoother.visualization.cluster_and_plot_celltype_props(p_inf_list, coords, names=None, n_col=4, transform='pca', n_neighbors=15, res=1, return_clu=False)

Cluster the cell-type proportions and visualize the results.

Parameters:

  • p_inf_list – list of cell-type proportions.

  • coords – n_spots x 2. Coordinates of the spots.

  • names – list of str. Names of the deconvolution model.

  • transform – str. Transformation to apply to the cell-type proportions. ‘pca’, ‘clr’, ‘ilr’.

  • n_neighbors – int. Number of neighbors to use for clustering.

  • res – float. Resolution for leiden clustering.

  • return_clu – bool. Whether to return the clustering results.